摘要
随着空间通信技术蓬勃发展,天地一体化信息网络不断推进,卫星通信网络资源管控愈发复杂。由于卫星资源稀缺、资源调度滞后于资源状态信息更新、业务不均匀分布等限制因素,如何进行高效资源管控成为卫星通信发展中亟待解决的问题之一。针对高低轨卫星异构网络系统架构,分析其网络资源管控面临的挑战。在综合分析传统管控架构的基础上,总结出基于分组管理的协同管控架构;针对卫星资源稀缺且分散的现状,阐述卫星网络虚拟资源池管理策略;引入基于深度强化学习的资源调度算法,解决传统调度方法在复杂环境下的失配问题;采用跳波束技术应对业务分布时空二维不均问题。
With the vigorous development of space communication technology and the continuous advancement of space-integrated-ground information network,satellite communication networks resource management and control is becoming more and more complex.Due to the scarcity of satellite resources,the slowness of resource scheduling compared to the status refreshing,and uneven distribution of business,effi ciently managing resources has become one of urgent problems to be solved in the development of satellite communications.In view of the heterogeneous network system architecture of high and low orbit satellites,the challenges,its network resource management and control facing,were analyzed.Integrating on the basis of traditional management and control architecture,the collaborative management and control architecture based on group management was introduced.The management strategy of satellite network virtual resource pool was explained to relieved resources scarcity.Resource scheduling algorithms based on deep reinforcement learning(DRL)was introduced to solved the mismatch problem of traditional scheduling methods in complex environments.Beam-hopping technology was adopted to deal with the two-dimensional unevenness of service distribution in time and space.
作者
张美蓉
镐梦婷
王闯
张更新
ZHANG Meirong;HAO Mengting;WANG Chuang;ZHANG Gengxin(College of Telecommunications&Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;A Certain Department of PLA,212000,China)
出处
《天地一体化信息网络》
2021年第4期66-74,共9页
Space-Integrated-Ground Information Networks
基金
国家自然科学基金资助项目(No.91738201)。
关键词
异构网络
资源管控
虚拟资源池
深度强化学习
跳波束技术
heterogeneous network
resource management and control
virtual resource pool
deep reinforcement learning
beam-hopping technology
